OpenAI promotes the adoption of GPT, third-party applications based on AI models, by allowing ChatGPT users to invoke them in any chat.
Starting today, paid users of ChatGPT, OpenAI’s AI chatbot interface, can include GPT in a conversation by typing “@” and selecting GPT from the list. The GPT you choose will understand the entire conversation, and different GPTs can be “tagged” for different exploit cases and needs – bringing the conversation into context with things that have been said previously.
“This allows you to add appropriate GPTs with full conversation context,” OpenAI said in a tweet.
You can now add GPT to any conversation in ChatGPT – just type @ and select GPT.
This allows you to add the appropriate GPTs with the full context of the conversation. pic.twitter.com/Pjn5uIy9NF
— OpenAI (@OpenAI) January 30, 2024
The move to boost GPT discoverability comes weeks after the launch of GPT Store, a GPT marketplace accessible through the ChatGPT dashboard. Creating GPTs requires no coding experience, and GPTs can be as straightforward or convoluted as the developer desires. Some of them are available today: trail recommender with AllTrails, coding teacher with Khan Academy, and content designer with Canva.
OpenAI plans to eventually introduce monetization for developers who want to sell access to their GPTs. However, the company may need to boost traffic first. According to data from Likeweb, a web analytics company, custom GPT tags so far only account for about 2.7% of global ChatGPT web traffic — and custom GPT traffic has been dwindling month over month since November.
Moderation proves to be another challenge. In the first week of launch, the GPT store was flooded with “romantic” chatbot apps, some of which had sexual overtones, in clear violation of OpenAI’s terms. Developers have also rushed to create political campaign bots – such as a chatbot impersonating US presidential candidate Dean Phillips – as another apparent violation.
OpenAI, which says it uses a combination of manual and automatic verification to flag GPT, has since removed some violating apps. However, if GPT volume increases, as the company is clearly counting on, it is conceivable that the problem will become even more acute.